PHILOSOPHY 4360/5360 – METAPHYSICS McTaggart’s Argument for the Unreality of Time I shall set out a response to McTaggart’s argument (a) that is very different from other responses, (b) that is clear-cut, but (c) that needs to withstand at least three important objections.

1. McTaggart's Argument McTaggart’s two-part argument against the reality of time: Part 1: There would be no time if nothing changed. Change cannot, contrary to Russell, be analyzed as a matter of an entity's having different properties at different times: the world involves genuine change only if events undergo change. But how can events change? McTaggart contends, first, that events cannot come into existence or drop out of existence, and, secondly, that they cannot change with respect to tenseless properties. Accordingly, the only way in which they can change is with respect to tensed properties. One response to part 1 of McTaggart’s argument: Genuine change, perdurantism, and causation. Part 2: Tensed properties are mutually incompatible. To avoid a contradiction, one must say that an event has the incompatible tensed properties at different times. But any specification of the relevant times will involve more complex tensed sentences, and those more complex tensed sentences will also have different truth-values at different times. So, once again, one needs to specify the relevant times. The result is a sequence of move and countermove that can then be repeated indefinitely, giving rise to an infinite regress. But this regress is, McTaggart holds, vicious, since one is left with a contradiction at every stage.

2. My Response to McTaggart's Argument My response is that there is no way of justifying the demand that the time at which something has a given tensed attribute be specified in tensed terms. Why should it matter whether the time is specified in a tensed fashion, or a tenseless one? For as long as the relevant times are specified, whether tensed or tenselessly, no contradiction will arise. But, if that is right, then the restriction to a tensed specification is unjustified, and, once that restriction is dropped, the argument collapses – for there is no difficulty in specifying in tenseless terms the times at which an event has various tensed attributes. Consider, for example, the event that is the birth of David Hume. It had, at different times, the properties of pastness, presentness, and futurity. If asked when it had those properties, there is no need to use tensed verbs to specify the times in question. One can, instead, use tenseless verbs plus dates. Thus one can say, for example, (1) The birth of David Hume is (tenselessly) future in 1710. (2) The birth of David Hume is (tenselessly) present at some specific moment in 1711. (3) The birth of David Hume is (tenselessly) past in 1712.

2

3. Possible Objections to this Response to McTaggart's Argument? What reasons might be offered for rejecting this way of responding to McTaggart’s argument? One objection that would be offered both by almost all advocates of tenseless approaches to time, and also by almost all advocates of tensed approaches to time, is that sentences (1), (2), and (3), though they contain the words “past”, “present”, and “future”, cannot really specify when the birth of David Hume has different tensed properties. For to do that, sentences (1), (2), and (3) would have to be tensed sentences, which they are not, since they are analytically equivalent, respectively, to the following tenseless sentences: (1*) The birth of David Hume is (tenselessly) later than 1710. (2*) The birth of David Hume is (tenselessly) simultaneous with some specific moment in 1711. (3*) The birth of David Hume is (tenselessly) earlier than 1712. A second objection, and one that some advocates of tensed approaches to time would certainly offer, is directed against the use of tenseless verbs in sentences (1), (2), and (3). For many advocates of tensed approaches to time claim that it is impossible to make any sense of such verbs. A third objection is that even if it is granted, for the sake of argument, that one can make sense of tenseless verbs, sentences (1), (2), and (3) do not really specify, in tenseless terms, when the birth of David Hume has the different tensed properties. The reason is that dates have to be analyzed in terms of temporal relations such as that of temporal priority, and temporal priority, in turn, has to be analyzed in terms of the tensed concepts of pastness, presentness, and futurity. So sentences (1), (2), and (3) turn out, upon scrutiny, to involve a tensed specification of the times in question.

4. Response to the First Objection Are the following two sentences analytically equivalent? (2) The birth of David Hume is (tenselessly) present at some specific moment in 1711. (2*) The birth of David Hume is (tenselessly) simultaneous with some specific moment in 1711. Consider the following tensed sentence: (2a) The birth of David Hume lies in the present. Does this sentence express the same proposition, or different propositions, when uttered at different times? Let time t be the moment in 1711 when David Hume was born. If someone uttered sentence (2a) at time t, the sentence would have been true. But if someone had uttered sentence (2a) in 1712, the sentence would have been false. Given that that’s so, isn’t it reasonable to conclude that utterances of the sentence “The birth of David Hume lies in the present” that are made at different times express different propositions?

3 Some advocates of tensed approaches to time – for example, Pavel Tichy1 – have maintained that this conclusion is mistaken, and that all tokens of a given tensed sentence express the same proposition. One way of attempting to support this view is by appealing to the distinction between truth simpliciter and truth at a time. But there are excellent reasons for rejecting the view in question. 1. Propositions that are true at one time cannot be false at a later time. 2. Tensed sentences that are inconsistent when uttered at the same time need not be inconsistent when uttered at different times. 3. Propositional attitude states involving tensed sentences. If tensed sentences express different propositions when uttered at different times, surely this can only be because such sentences contain an indexical, and an indexical, moreover, that points, directly or indirectly, to a time – the time, namely, when the utterance is made. Is such indexicality all there is to tensed sentences? If it is, then sentence (2a) could be rewritten as (2b) The birth of David Hume occurs (tenselessly) at this very moment. Claim: Acceptance of the idea that ordinary tensed sentences contain indexicals referring, directly or indirectly, to the time of utterance does not entail, however, that a tenseless analysis of tensed utterances is correct. Compare (2b) with (2c) The birth of David Hume has (tenselessly) the property of presentness at this very moment. Given a sentence containing an indexical, one can always replace the indexical with a name, or a definite description, that refers to the same thing, and the resulting sentence, while different in meaning, will express the same general sort of fact. Doing this with sentence (2c) for example, might give one (2d) The birth of David Hume has (tenselessly) the property of presentness at time t. My response to the first objection is, in short, as follows: (a) Ordinary tensed sentences express different propositions when uttered at different times. (b) This is because an ordinary tensed sentence contains an indexical term that points, directly or indirectly, at a time – namely, the time of utterance. (c) If an ordinary tensed sentence functions to attribute a tensed property to an event, the sentence that results when one replaces the indexical referring to a time by a date referring to that very same time will equally function to attribute the same tensed property to the same event. (d) The conclusion, accordingly, is that there is no ground for holding that sentences (1), (2), and (3) are not tensed sentences.

1

70.

Pavel Tichy, “The Transiency of Truth,” Theoria, 46 (1980), 165-82. See pp. 166-

4 A shorter route to the same conclusion is to point out that one way of interpreting sentences (1), (2), and (3) is as analytically equivalent to (1**) The birth of David Hume has (tenselessly) the tensed property of futurity in 1710. (2**) The birth of David Hume has (tenselessly) the tensed property of presentness at some moment in 1711. (3**) The birth of David Hume has (tenselessly) the tensed property of pastness in 1712.

5. Response to the Second Objection The second objection is that my response to McTaggart’s argument makes use of tenseless verbs, and that it is impossible to make any sense of such verbs. 1. This view would entail that tenseless approaches to time are unintelligible. 2. The appeal to mathematical statements: 2 + 2 is equal to 4. 3. The case of instantaneous events having intrinsic properties, or standing in relations. 4. The case of instantaneous temporal parts of persisting entities having intrinsic properties, or standing in relations. 5. Sentences constructed from terms for persisting object, predicates that attribute intrinsic properties, and time terms. Conclusion: The idea of there being sentences that contain no tensed concepts, but that nevertheless express propositions – including propositions about temporal states of affairs – does not seem to me at all problematic.

6. Response to the Third Objection The third objection was that sentences (1), (2), and (3) involve dates, and dates have to be analyzed in terms of temporal relations such as that of temporal priority, which, in turn, has to be analyzed in terms of the tensed concepts of pastness, presentness, and futurity. So sentences (1), (2), and (3) turn out, upon scrutiny, to involve a tensed specification of the times in question. Does this objection make it out of the starting blocks? A reason for thinking it does not is that McTaggart’s argument makes use of tensed sentences, tokens of which have different truth-values at different times, and this seems absolutely crucial to his argument. But sentences (1), (2), and (3), regardless of how they are analyzed, do not share this feature. My fundamental response to this third objection, however, is simply that it is impossible to analyze temporal priority in terms of tensed concepts. (1) If the concept of temporal priority is to be analyzed in tensed terms, one will need to make use, at a minimum, of the concepts of pastness, presentness, and futurity. (2) One needs to ask whether it is being claimed that all three of those concepts are analytically basic. (3) If it is, then it can be argued that this is false in the case of the concepts of pastness and futurity.

5 The reason is that there must be a criterion that distinguishes between concepts that are analytically basic and those that are not, and in the case of descriptive concepts – as contrasted with logical concepts – the two most plausible criteria are as follows: The Direct Acquaintance Criterion Concept C can be analytically basic for a person P only if it is the concept of a property or relation such that P has been directly acquainted with instances of that property or relation. The Immediate Perception Criterion Concept C can be analytically basic for a person P only if it is the concept of a property or relation such that P has immediately perceived instances of that property or relation. The claim is then that human persons have not immediately perceived events having the property of futurity, let alone been directly acquainted with the property of futurity. So the concept of futurity cannot be analytically basic. The question, accordingly, is how the concept of lying in the future is to be analyzed. (1) An analysis in terms of the later than relation? (2) An analysis in terms of causation? To sum up, the situation is that the concept of lying in the future is not analytically basic, and when one attempts to analyze that concept, one has to employ concepts that, on their own, and without any tensed concepts, will allow one to define the earlier than relation. Accordingly, sentences (1), (2), and (3), involving the concept of dates, and thereby, concepts of temporal relations, including temporal priority, do not involve any tensed concepts. For any proposed analysis of, for example, the relation of temporal priority in tensed terms will either involve a concept – such as that of lying in the future – that is falsely being claimed to be analytically basic, or else will involve non-tensed concepts that are, in themselves, sufficient to provide an analysis of temporal priority, so that the proposed tensed analysis is merely adding on tensed concepts that are unnecessary to a simpler, and perfectly satisfactory, tenseless analysis, such as one in terms of causation.

7. Summing Up If one can specify the times at which an event has the different tensed properties, not in tensed terms, but tenselessly, by means of sentences that combine tenseless verbs plus dates, then one has a clear cut refutation of McTaggart’s argument. This way of proceeding is open, however, to certain objections. I have argued, however, that what I take to be the three crucial objections are unsound.